Carbon molecular sieve membranes for nitrogen/methane separation

Technical field

The various embodiments of the disclosure relate generally to carbon molecular sieve membranes (CMSM) and their associated fabrication processes, and more particularly to CMSM that separate nitrogen and methane.

Background

Many natural gas reserves are highly contaminated by nitrogen, which has to be removed to meet pipeline specifications (nitrogen below 3%). Nitrogen rejection technology is also required in some chemical refineries where methane stream has to be purified before certain processes like steam reforming. Nitrogen removal from natural gas is a difficult and expensive process. It is estimated that nitrogen removal is more challenging and costly than corresponding carbon dioxide removal from coal bed methane. Currently there are four available nitrogen rejection technologies—cryogenic distillation, pressure swing adsorption (PSA), liquid-based absorption, and membrane separation (CH.sub.4-selective). Among them, only cryogenic distillation is a mature technology that has been commercially used on large scale. Although the cryogenic process has the highest methane recovery rate (about 98%), the complexity of pretreatment and high energy consumption make it less attractive for small throughputs. PSA and liquid absorption processes are better suited for small throughputs (<10 MMscfd) with relatively high nitrogen concentration. But both processes are capital intensive. Membrane technology is an attractive alternative compared to other separation technologies by virtue of its high adaptability, high reliability, low energy consumption, and low capital cost in installation, operation and maintenance. It utilizes the “free” well pressure as the driving force and separates enriched fast and slow gases into permeate and retentate streams respectively. With its modular nature and good space efficiency, membrane gas separation is especially attractive to small and remote sites.

The future growth of natural gas consumption and depletion of conventional resources will promote the development of low quality natural gas (LQNG) reserves, which account for almost one-third of known gas reserves in the US. Such LQNG contains excessive amounts of gas other than methane, mostly nitrogen, carbon dioxide, and hydrogen sulfide. It has been estimated that approximately 17% of the natural gas reserves in the US are sub-quality due to high levels of nitrogen (typically 5% to 40%). Moreover, considerable amount of nitrogen-rich natural gas (NRNG) reserves (with 10 to 85% N.sub.2) exist in many different regions in the world. Apart from the natural existence, nitrogen contamination can also be artificially introduced by nitrogen fracturing during some shale gas production processes. In such case, nitrogen content can reach as high as 50% at the start of work-up and then decline to 6-10%. The NRNG needs to be upgraded before it is sent to the commercial natural gas pipeline. The US natural gas pipeline specification typically limits nitrogen to be below 3%.

Brief summary

The various embodiments of the disclosure relate generally to carbon molecular sieve membranes and their associated fabrication processes, and more particularly to CMSM that separate nitrogen from methane.

An embodiment of the disclosure can be a gas separation membrane of nitrogenmethane separation. The gas separation membrane can include a carbon molecular sieve. The gas separation membrane has a nitrogenmethane selectivity of at least about 5.0, and a nitrogen permeability of at least about 3.0 Barrers at 35° C. and 65 psia for pure nitrogen and pure methane streams. In some embodiments, the carbon molecular sieve can be a pyrolyzed polymer precursor membrane of a polyimide having at least about 60 wt % carbon content and a hydrogen content of less than about 5% wt.

In some embodiments, the gas separation membrane can have a nitrogenmethane selectivity of at least about 5.5 or at least about 6. In some embodiments, the gas separation membrane can have a nitrogen permeability that is at least about 3.9 Barrers, at least about 4 Barrers, or at least about 5 Barrers.

In some embodiments, the gas separation membrane can include the carbon molecular sieve of a pyrolyzed polymer precursor membrane selected from

##str00001##

Another embodiment of the disclosure can be a process for forming a nitrogenmethane separation carbon membrane. The process can include subjecting a polymer precursor membrane to a pyrolysis temperature of at least about 675° C. in an inert atmosphere. The polymer precursor membrane can be a polyimide having at least about 60 wt % carbon content and a hydrogen content of less than about 5% wt. In some embodiments, the pyrolysis temperature can be at least about 700° C., or at least about 750° C., or at least about 800° C. In other embodiments, the polymer precursor membrane does not contain fluorine. In an embodiment, the inert atmosphere for pyrolysis is argon, or ultrahigh purity argon.

In an embodiment, the process can produce a carbon membrane having a nitrogenmethane selectivity of at least 5, or a nitrogenmethane selectivity of at least 6. In some embodiments, the process can product a carbon membrane having a nitrogen permeability that is at least about 3.9 Barrers, at least about 4 Barrers, or at least about 5 Barrers. In some embodiments, the polymer precursor membrane in the pyrolysis can be a polyimide selected from

##STR00002## or any combination thereof, as a polymer blend or copolymer.

An embodiment of the disclosure can be a gas enrichment device, the device having a nitrogenmethane gas stream inlet, a nitrogen-enriched gas stream outlet, a nitrogen-depleted gas stream outlet, and a carbon molecular sieve membrane interposed between the gas stream inlet and the nitrogen-depleted stream outlet on one side of the membrane and the nitrogen-enriched gas stream outlet on the other side. The carbon molecular sieve can have a nitrogenmethane selectivity of at least about 5.0 and a nitrogen permeability of at least about 3 Barrers at 35° C. and 65 psia in pure nitrogen and pure methane streams. The carbon molecular sieve can be a pyrolyzed polymer precursor membrane of a polyimide having at least about 60 wt % carbon content and a hydrogen content of less than about 5% wt. The nitrogenmethane selectivity can be at least about 5.5, or at least about 6, and the nitrogen permeability can be at least about 3.9 Barrers, at least about 4 Barrers, or at least about 5 Barrers.

An embodiment of the disclosure can also be a method for decreasing the nitrogen content of a nitrogenmethane gas mixture. The method can include contacting the mixture with a gas separation membrane, and separating two gas streams at the membrane, wherein the gas separation membrane is as described above.

Brief description of the drawings

FIG. 1 illustrates structures of polyimides in carbon molecular sieve (CMS) membrane fabrication, in accordance with an exemplary embodiment of the disclosure.

FIG. 2 illustrates N.sub.2 /CH.sub.4 Separation performance of CMS membranes (with legend on right) plotted with data of polymer membranes from literature (red circles) and the Robeson line, in accordance with exemplary embodiments of the disclosure.

FIGS. 3( a ) and 3( b ) illustrate structure of pyrolytic carbon, FIG. 3 ( c ) illustrates idealized pore structure of CMS membrane, and FIG. 3( d ) illustrates idealized bimodal pore size distribution of CMS membrane, with the notations correspond to d.sub.c: ultramicropore dimension, d.sub.tv: size of adsorptive pore dimension, and d.sub.λ: jump length dimension, in accordance with exemplary embodiments of the disclosure.

FIG. 4 illustrates a schematic of a pyrolysis system, in accordance with exemplary embodiments of the disclosure.

FIG. 5 illustrates N.sub.2 permeability and N.sub.2 /CH.sub.4 permselectivity of Matrimid® precursor and CMS dense films pyrolyzed at 550, 675, and 800° C. Each CMS data points represent an average of two CMS films with the error bars shown (the error bars for 800° C. CMS are too small to be seen) in accordance with exemplary embodiments of the disclosure.

FIG. 6 illustrates idealized hypothetical ultramicropore size distribution for CMS membranes pyrolyzed at three final pyrolysis temperatures, in accordance with embodiments of the disclosure.

FIG. 7 illustrates temperature dependence of permeabilities of 800° C. CMS membrane at 65 psia (permeability P is in the unit of Barrer) in accordance with exemplary embodiments of the disclosure.

FIG. 8 illustrates N.sub.2 and CH.sub.4 Langmuir isotherms of 800° C. CMS dense film at 25, 35 and 50° C. in accordance with exemplary embodiments of the disclosure.

FIG. 9 illustrates temperature dependence of sorption coefficients of a 800° C. CMS membrane at 65 psia (sorption coefficient is in the unit of cm.sup.3 (STP)/cm.sup.3 CMS/psi) in accordance with exemplary embodiments of the disclosure.

FIG. 10 illustrates temperature dependence of diffusion coefficients of a 800° C. CMS membrane at 65 psia (diffusion coefficient D is in the unit of 10.sup.−8 cm.sup.2/s) in accordance with exemplary embodiments of the disclosure.

FIG. 11 illustrates sizes and shapes of N.sub.2 and CH.sub.4 molecule in accordance with exemplary embodiments of the disclosure.

FIG. 12( a )-( d ) illustrate configurations of N.sub.2 and CH.sub.4 in the normal and activated state, (a) top view of N.sub.2 and CH.sub.4 in adsorbed normal state, (b) side view of N.sub.2 and CH.sub.4 in adsorbed normal state, (c) side view of N.sub.2 and CH.sub.4 in activated transition state during a diffusion jump, in accordance with exemplary embodiments of the disclosure.

FIG. 13 illustrates a gas separation device in accordance with exemplary embodiments of the disclosure.

FIG. 14 illustrates a gas separation device in accordance with exemplary embodiments of the disclosure.

Detailed description

Although preferred embodiments of the disclosure are explained in detail, it is to be understood that other embodiments are contemplated. Accordingly, it is not intended that the disclosure is limited in its scope to the details of construction and arrangement of components set forth in the following description or illustrated in the drawings. The disclosure is capable of other embodiments and of being practiced or carried out in various ways. Also, in describing the preferred embodiments, specific terminology will be resorted to for the sake of clarity.

It must also be noted that, as used in the specification and the appended claims, the singular forms “a,” “an” and “the” include plural referents unless the context clearly dictates otherwise.

Also, in describing the preferred embodiments, terminology will be resorted to for the sake of clarity. It is intended that each term contemplates its broadest meaning as understood by those skilled in the art and includes all technical equivalents which operate in a similar manner to accomplish a similar purpose.

Ranges may be expressed herein as from “about” or “approximately” one particular value andor to “about” or “approximately” another particular value. When such a range is expressed, another embodiment includes from the one particular value andor to the other particular value.

By “comprising” or “comprising” or “including” is meant that at least the named compound, element, particle, or method step is present in the composition or article or method, but does not exclude the presence of other compounds, materials, particles, method steps, even if the other such compounds, material, particles, method steps have the same function as what is named.

It is also to be understood that the mention of one or more method steps does not preclude the presence of additional method steps or intervening method steps between those steps expressly identified. Similarly, it is also to be understood that the mention of one or more components in a device or system does not preclude the presence of additional components or intervening components between those components expressly identified.

Membranes with high selectivity for separating nitrogen and methane have been developed that provide substantial improvements in selectivity and permeability compared to previous efforts. This disclosure describes the CMS membranes for nitrogenmethane separation, methods of preparing the CMS membranes, methods of using the CMS membranes, and devices that incorporate the CMS membranes.

Gas transport through non-porous polymers and molecular sieves is commonly described via the “solution-diffusion” mechanism. Gas molecules sorb into the membrane at the high pressure upstream, then diffuse through the membrane under the influence of a chemical potential gradient, and finally desorb from the membrane at the low pressure downstream side. With this mechanism, permeability of a membrane for a certain gas can be expressed as a product of a thermodynamic factor called sorption coefficient, S, and a kinetic factor called the diffusion coefficient, D. And the permselectivity is a product of sorption selectivity and diffusion selectivity. For all membrane materials that have been reported, the sorption coefficient of CH.sub.4 is always larger than that of N.sub.2 due to the higher condensability of CH.sub.4. Thus sorption effect enables CH.sub.4 to permeate faster. However, the diffusion coefficient of N.sub.2 is larger than that of CH.sub.4 in all kinds of membrane materials due to the smaller kinetic diameter of N.sub.2, which means diffusion coefficient enables N.sub.2 to permeate faster. With the effects of sorption and diffusion opposed to each other, there are two types of membranes for N.sub.2/CH.sub.4 separation—N.sub.2-selective membranes with stronger diffusion selection effect and CH.sub.4-selective membranes with stronger sorption selection effect. For rubbery polymer materials, a broad distribution of segmental motions produces a correspondingly broad distribution of transient intersegmental gap sizes and poor distinguishability of gas molecule sizes and shapes. The membranes fabricated from them are usually CH.sub.4-selective. For glassy polymers, limited mobility of polymer segments leads to a relatively narrow distribution of effective transient intersegmental gap sizes and better distinguishability of gas molecule sizes and shapes. The membranes fabricated from them are usually N.sub.2-selective. MTR Inc. has been developing CH.sub.4-selective rubbery polymer membrane systems for N.sub.2 removal from natural gas. They have achieved a CH.sub.4/N.sub.2 selectivity of 3 at ambient temperature. This disclosure focuses on N.sub.2-selective membranes which attempt to maximize the diffusion-selection effect. With N.sub.2 passing through the membrane and CH.sub.4 rejected, the wellhead pressure of CH.sub.4 can be largely preserved and the cost of recompression in a CH.sub.4-selective membrane system can be saved.

The technologies of nitrogenmethane separation traditionally included cryogenic distillation, liquid absorption, solid adsorption, and methane-selective membrane separation. In contrast, this disclosure describes a class of nitrogen-selective membranes to remove nitrogen from methane. Compared to other technologies, membrane separation has lower energy consumption and lower capital cost in installation, operation and maintenance. Compared to methane-selective membranes, nitrogen-selective membranes described in this work can largely preserve the wellhead pressure of methane, since nitrogen permeates through the membrane and methane is rejected at the retentate side of the membrane. Thus the cost of recompression of purified methane stream can be saved.

Polymer-based N.sub.2-selective membrane separation for N.sub.2/CH.sub.4 has not been pursued widely. This is because that the broad distribution of segmental motions of polymer chains leads to low permselectivity for N.sub.2/CH.sub.4. Also, there is a strong trade-off effect between productivity (represented by permeability) and efficiency (represented by selectivity) as shown in the upper bound line for N.sub.2/CH.sub.4 drawn by Robeson (see FIG. 2 ). The separation performance of all traditional polymer membranes lies either below or on the upper bound line.

Membrane technology is an attractive alternative to other separation technologies by virtue of its potentially high adaptability, low energy consumption, and low capital cost. With its modular nature and good space efficiency, membrane based gas separation is especially attractive to small and remote sites; however, the competing effect between diffusion and sorption terms poses a challenge for membrane technology to separate N.sub.2 and CH.sub.4. Nitrogen is smaller (kinetic diameters: 3.64 Å vs. 3.8 Å) but less condensable than CH.sub.4 (critical temperatures: 126 K vs. 191 K). These properties make N.sub.2 permeate faster than CH.sub.4 in membranes with effective diffusion selection abilities, while CH.sub.4 permeates faster in membranes with stronger sorption selection effect. Most rubbery polymeric membranes are slightly CH.sub.4-selective, since flexible chains in rubbery polymer provide very small N.sub.2/CH.sub.4 diffusion selectivity; therefore CH.sub.4/N.sub.2 solubility selectivity dominates permselectivity. The highest CH.sub.4/N.sub.2 permselectivity of CH.sub.4-selective membranes has been achieved around 3˜4 at ambient temperature. Although the flux is high for this membrane system, loss of high well-head pressure of the permeate requires additional costs for recompression. In 1987, Kim and Koros et al. identified polyimides with “reverse selectivity”, i.e., N.sub.2 permeates preferentially rather than CH.sub.4. With CH.sub.4 rejected to the high pressure retentate stream, the advantage of saving recompression cost is significant. Unfortunately, the best performing N.sub.2-selective glassy polymer membranes with acceptable N.sub.2 permeability only provide a N.sub.2/CH.sub.4 permselectivity ˜3, which is marginal for economical separations. Although glassy polyimide membranes have limited mobility of polymer segments, the diffusion selection effect is insufficient to distinguish the small size and shape difference between N.sub.2 and CH.sub.4. As a result, a more rigid membrane structure with appropriate pore size and shape is required to fully utilize the size and shape difference between N.sub.2 and CH.sub.4 to provide the required diffusion selectivity and overcome the higher sorption of CH.sub.4 vs. N.sub.2.

Carbon molecular sieve (CMS) membranes, pyrolyzed from polymer precursors under controlled conditions have shown attractive separation performance for several gas pairs. CMS hollow fiber membranes have also been proved stable for high pressure feeds up to 1000 psi, (1 psi=0/006895 bar) without showing plasticization that are commonly encountered in polymer membranes. However, only a few researchers have reported N.sub.2 permeation data of CMS membranes, and most of them did not focus on N.sub.2 /CH.sub.4 separation, but rather on O.sub.2/N.sub.2 and CO.sub.2/CH.sub.4 separation. The reported data for these CMS membranes showed N.sub.2/CH.sub.4 permselectivity in the range of 1 to 4 but two cases had higher permselectivity. Steel et al. prepared Matrimid® derived CMS membrane at 800° C. under vacuum and achieved N.sub.2/CH.sub.4 permselectivity of 5.89 with N.sub.2 permeability of 1.86 Barrers. Hosseini et al. fabricated CMS membrane pyrolyzed from PBIMatrimid® (25/75 wt. %) blend polymers (vacuum, 800° C.) and achieved N.sub.2 /CH.sub.4 permselectivity of 7.99 with N.sub.2 permeability of 3.78 Barrers.

Several polyimide polymers as the precursors for CMS membranes are described herein. Polyimides are chosen due to their outstanding properties such as high glass transition temperatures, good processability, good mechanical strength, high carbon yields, and low shrinkage. The polyimides can be pyrolyzed under inert atmosphere, including under ultra high purity (UHP) argon, with temperature gradually increased to 800° C. and soaked at this final temperature for 2 hr. The high pyrolysis temperature (above 675° C.) is critical to achieve high N.sub.2 /CH.sub.4 selectivity. Higher pyrolysis temperature usually leads to lower permeability and higher selectivity of CMS membranes, because graphene-like sheets begin to align parallel on a short range and pack more closely to each other as temperature increases. The difference of kinetic diameters of N.sub.2 and CH.sub.4 is very small (3.64 Å vs. 3.80 Å). Thus increasing pyrolysis temperatures helps to decrease the size of ultramicropores to approach the sizes of gas molecules and highly improve the molecular sieving effect. Apart from the small size difference, the counter effect of sorption competition is another challenge for the membrane separation of N.sub.2 /CH.sub.4. The sorption coefficient of CH.sub.4 is much higher than that of N.sub.2 due to the higher condensability of CH.sub.4. Thus the diffusion selectivity has to be high enough to overcome the low sorption selectivity to achieve a satisfying overall permselectivity.

In an embodiment, CMS membranes can be used as a gas separation membrane for separating nitrogen and methane. The gas separation membrane can contain a CMS with a nitrogenmethane selectivity of at least about 5.0, and a nitrogen permeability of at least about 3.0 Barrers. The nitrogenmethane selectivity can be at least about 5.5, at least about 6.0, at least about 6.5, or at least about 7.0. The nitrogen permeability can be at least 3.5 Barrers, at least about 3.9 Barrers, at least about 4.0 Barrers, at least about 5.0 Barrers, at least about 5.5 Barrers, or at least about 6.0 Barrers. In some embodiments, the nitrogen permeability can be at least about 3.5 Barrers and the nitrogenmethane selectivity at least about 6.0, the nitrogen permeability can be at least about 5.0 Barrers and the nitrogenmethane selectivity at least about 5.5, the nitrogen permeability can be at least about 6.0 Barrers and the nitrogen/methane selectivity at least about 7.0, or the nitrogen permeability can be at least about 6.0 Barrers and the nitrogenmethane selectivity at least about 7.5. The nitrogen permeability can be up to about 50 Barrers and the nitrogen/methane selectivity can be up to about 30.

Values are measured at 35° C. and 65 psia for pure nitrogen and pure methane streams.

While the need for nitrogenmethane separation is highly desirable, and current results have achieved the surprising ranges above, early efforts to create CMS membranes that separated nitrogen and methane have failed to produce both a highly selective membrane that also exhibited acceptable permeability. Early efforts focused on materials that had shown excellent results in other gas pairings, including for example the 6FDA imides, structures VI-VIII, shown in FIG. 1 . However, a variety of efforts to develop these previously successful materials did not lead to a suitable material, often yielding good to excellent permeability, but almost no selectivity for nitrogen and methane. Similar efforts to apply a dual temperature secondary oxygen doping strategy previously described by this group also did not lead to a suitable material. Surprisingly, by switching to a more traditional material and optionally avoiding materials that exhibited fluorine groups, the pyrolysis could be tailored to produce a product that achieves selectivity and permeability never before demonstrated.

Thus, in an embodiment, the carbon molecular sieve membrane can be a pyrolyzed polymer precursor membrane. The polymer precursor membrane can be a polyimide, preferably a polyimide with a high fraction of carbon atoms are sp.sup.2 hybridized atoms, or more specifically an aromatic polyimide wherein the portion of atoms are conjugated with an aromatic ring. In an embodiment, the fraction of carbon atoms that are sp.sup.2 hybridized is at least about 70 atom %, at least about 75 atom %, or at least about 80 atom %. In another embodiment, the portion of atoms conjugated to an aromatic ring can be at least about 45 atom %, or at least about 50 atom %.

In another embodiment, the polyimide can be described based on the amount of carbon and hydrogen present as its molecular weight. In another embodiment, the polyimide can be at least about 60 wt % carbon and less than about 5 wt % hydrogen, at least about 65% weight carbon and less than about 5 wt % hydrogen, or at least about 70 wt % carbon and less than about 5 wt % hydrogen.

By way of a nonlimiting example, the polyimide Matrimid, structure III in FIG. 1 , has a molecular formula of approximately C.sub.35H.sub.24N.sub.2O.sub.5. The compound has about 76.1 wt % carbon and about 4.3 wt % hydrogen. About 82% of the carbons (29 out of 35) are sp.sup.2 hybridized. About 54% of the atoms (29 carbons, 2 nitrogens, and 5 oxygens) are conjugated to, i.e. in it election conjugation, with an aromatic ring.

Based on work conducted using the previous 6FDA compounds, it also appears that the trifluoro methyl groups surprisingly change the reactivity of polyimides during pyrolysis in such a way that the final CMS membrane does not achieve the characteristics desired by the disclosure. Thus, in an embodiment, the polymer precursor does not contain a fluorine group in the structure.

In some embodiments, the polymer precursor membrane can be a polyimide selected from the group consisting of structures I-V, shown below

##STR00003## In some preferred embodiments, the polyimide can be selected from the structures I, II, III or IV. In some embodiments, the polyimide can be selected from the structures III or IV. Moreover, while structures I-V are displayed as single polymer compositions, the polyimide used to produce that CMS membranes can also include combinations of the polyimides, as blends of two or more polyimides, or as a copolymer of two or more polyimides.

Alternatively, the polyimide can be prepared from an diacid anhydride and a aromatic diamine, such that the reaction would produce a polyimide having a structure that includes the imide IX or X, shown below:

##str00004##

As discussed above, CMS membranes can be pyrolyzed from polymer membrane precursors under controlled atmosphere and temperature protocol. Without wishing to be bound by theory, during pyrolysis, scission and reorganization of polymer main chains occur, with small gas molecules released and free radicals formed. The free radicals then form interchain or intrachain crosslinking, giving rise to conjugated aromatic 3D network with graphene-like structure. As the pyrolysis temperature increases, sp.sup.2 hybridized carbon sheet begin to align to form ordered structures on a short range. On the long range, these turbostratic structures are believed to be amorphous and isotropic as illustrated in FIG. 3( a, b ) . In FIGS. 3( a ) and ( b ) are shown the structures of pyrolytic carbon, FIG. 3 ( c ) shows the idealized pore structure of CMS membrane, and FIG. 3 ( d ) shows the idealized bimodal pore size distribution of CMS membrane. In ( c ), the notations correspond to d.sub.c: ultramicropore dimension, d.sub.tv: size of adsorptive pore dimension, and d.sub.λ: jump length dimension. ‘Slit-like’ pore structures are formed from packing imperfections between ordered regions, and a bimodal pore distribution can be used to describe CMS structures—‘micropores’ with size of 7 to 20 Å connected by molecular sieving ‘ultramicropores’ with sizes below 7 Å. This morphology enables high permeability and high selectivity at the same time [26].

Transport in CMS membranes can be described in terms of permeability and perselectivity, or selectivity. ‘Permeability’ and ‘permselectivity’ are two intrinsic properties to characterize the intrinsic productivity and separation efficiency of membranes. The ‘permeability’ of a membrane to gas A is defined as the pressure-and-thickness-normalized flux, given as

P A = N A .Math. l Δ ⁢ ⁢ p A ( 1 ) where N.sub.A is molar flux of A, l is membrane thickness, and Δp.sub.A is the partial pressure difference between upstream and downstream for gas A. It is common to express permeability in Barrers, for which the units are given below.

1 ⁢ ⁢ Barrer = 10 - 10 ⁢ cm 2 ⁡ ( STP ) .Math. cm cm 2 .Math. s .Math. cm ⁢ ⁢ Hg ( 2 )

Gas transport through non-porous polymers and molecular sieves is commonly described via the ‘solution-diffusion’ mechanism. Gas molecules sorb into the membrane at the high pressure upstream, then diffuse through the membrane under the influence of a chemical potential gradient, and finally desorb from the membrane at the low pressure downstream side. Considering the sorption-diffusion nature of the process, permeability can be written as a product of a kinetic factor called the diffusion coefficient D.sub.A, and a thermodynamic factor called the sorption coefficient .sub.A. P .sub.A =D .sub.A.Math. .sub.A

Due to their microporous nature, the sorption coefficient in CMS membranes can be described by Langmuir isotherm, viz.

�� A = C A p A = C HA ′ ⁢ b A 1 + b A ⁢ p A ( 4 ) where C.sub.A is the concentration of sorbed penetrant A, p.sub.A is the partial pressure, C′.sub.HA is the Langmuir hole-filling capacity and b.sub.A is the Langmuir affinity constant. For polymer materials, gas penetrants may sorb either into well packed Henry's law regions by dilation of the matrix, or into the preexisting gaps between polymer chain segments comprising Langmuir regions [28]. Thus the sorption in polymer membranes can be described by the following equation,

�� A = C A p A = k DA + C HA ′ ⁢ b A 1 + b A ⁢ p A ( 5 ) where k.sub.DA is the Henry's law coefficient.

The separation factor between gases A and B is defined as the ratio of permeate-side and feed-side mole fractions of A and B. When downstream pressure is held at vacuum, the separation factor equals to the intrinsic permselectivity of the membrane, α.sub.A/B [27]. The permselectivity is also a product of diffusion selectivity

D A D B and sorption selectivity

�� A �� B , so:

α A / B = ( y A / y B ) permeate ( x A / x B ) feed = P A P B = D A D B .Math. �� A �� B ( 6 ) Gas diffusion in CMS membrane is an activated process, which requires penetrants to have enough activation energy to overcome repulsion from size discriminating pores, so D.sub.A follows an Arrhenius relationship.

D A = D 0 ⁢ ⁢ A ⁢ exp ⁡ ( - E DA RT ) ( 7 ) where R is the molar gas constant, T is the temperature, D.sub.0A and E.sub.DA are the pre-exponential factor and activation energy for diffusion of penetrant A, respectively. The thermodynamically-based sorption coefficient .sub.A decreases with temperature, following van't Hoff equation.

�� A = �� 0 ⁢ ⁢ A ⁢ exp ⁡ ( - H SA RT ) ( 8 ) where .sub.0A is the preexponential factor for sorption, and H.sub.SA is the apparent heat of sorption, which is typically negative. From Eqs. (3),

and (8), we get

0 P A = P 0 ⁢ ⁢ A ⁢ exp ⁡ ( - E PA RT ) ⁢ ⁢ where ( 9 ) P 0 ⁢ ⁢ A = D 0 ⁢ ⁢ A .Math. �� 0 ⁢ ⁢ A ( 10 ) E PA = E DA + H SA ( 11 )

Since the increase in diffusion coefficient typically exceeds the decrease in sorption coefficient with rising temperature, permeability usually increases with temperature.

As the rigid ultramicropore window size (d.sub.c in FIG. 3 ( c ) ) approaches the minimum dimension of gas penetrants, very effective size and shape diffusion selection becomes possible, thereby providing high diffusion selectivity. Unlike polymer membranes, highly rigid size and shape media such as CMS materials can introduce an additional factor, termed ‘entropic diffusion selectivity’. This factor can provide additional diffusion selectivity by differentiating the pre-exponential factor (D.sub.0A) in favor of penetrants that do not give up degrees of freedom in the diffusion transition state. As will be discussed later, the linear nature of N.sub.2 enables it to benefit from small slit-shaped pores typical of CMS. This simple feature will be shown to be useful in understanding the N.sub.2/CH.sub.4 selectivity of our CMS samples.

In order to achieve the described selectivity and permeability, the process for preparing the CMS membranes had to be developed that achieved the preferred structural aspects discussed above. In an embodiment, a process for forming a nitrogenmethane separation carbon membrane can include subjecting a polymer precursor membrane to a pyrolysis temperature of at least about 675° C. in an inert atmosphere, wherein the polymer precursor membrane is a polyimide as described above. In an embodiment, the pyrolysis temperature can be at least 700° C., at least about 750° C., or at least about 800° C. The polyimide can have a fraction of its carbon atoms that are sp.sup.2 hybridized of at least about 70 atom %, at least about 75 atom %, or at least about 80 atom %. The polyimide can have a portion of atoms conjugated to an aromatic ring can be at least about 45 atom %, or at least about 50 atom %. Or, the polyimide can be at least about 60 wt % carbon and less than about 5 wt % hydrogen, at least about 65% weight carbon and less than about 5 wt % hydrogen, or at least about 70 wt % carbon and less than about 5 wt % hydrogen.

In an embodiment, the pyrolysis can be conducted in an inert atmosphere. In an embodiment, the pyrolysis can be conducted in an argon atmosphere, preferably an ultrahigh purity argon atmosphere.

In an embodiment, the polymer precursor can be a polyimide selected from formulas I, II, III, IV, or V, preferably formulas I, II, III, or IV, or preferably formulas III or IV. In an embodiment, the polymer precursors can be blends of two or more of formulas I, II, III, IV, or V, or copolymers of two or more of formulas I, II, III, IV, or V.

CMS dense films were successfully prepared from commercial polyimide Matrimid® 5218 for N.sub.2 /CH.sub.4 separation. CMS films showed significant improvement in separation performance compared to the polymer precursor. It was found that N.sub.2 permeability decreases and N.sub.2 /CH.sub.4 permselectivity increases with the increase of pyrolysis temperature. Very attractive separation performance was observed for 800° C. CMS dense film. With N.sub.2 permeability of 6.78 Barrers and N.sub.2 /CH.sub.4 permselectivity of 7.69 in pure gas permeation and N.sub.2 permeability of 5.17 Barrers and N.sub.2 /CH.sub.4 permselectivity of 5.94 in mixed gas permeation test, the permeation performance of this sample exceeds the polymer upper bound line. This type of high-performing CMS material has also been shown processable for the hollow fiber morphology.

Upon further study of this 800° C. CMS dense film, it was found that the high permselectivity is solely contributed by diffusion selectivity. Temperature dependences of permeabilities, sorption coefficients and diffusion coefficients were studied for this high performing material. It was found that permeabilities and diffusion coefficients increase with temperature, following an Arrhenius relationship, and sorption coefficients decrease with temperature, following a van't Hoff relationship. Temperature dependences of sorption coefficients revealed that the heat of adsorption of N.sub.2 is slightly more negative than that of CH.sub.4. As a result, the diffusion jump requires higher activation energy for N.sub.2 than CH.sub.4. Though diffusion activation energy makes negative contribution to the N.sub.2 /CH.sub.4 diffusion selectivity, diffusion activation entropy compensates this effect and highly enhances the diffusion selectivity and permselectivity. This entropic diffusion selection effect is achieved by a very tight ‘slit-shaped’pore structure that can effectively distinguish the size and shape difference between N.sub.2 and CH.sub.4.

With the development of the new materials, several additional disclosures can be described in view of these nitrogen/methane CMS membranes. In an embodiment, the disclosure can include a method for decreasing the nitrogen content of a nitrogenmethane gas mixture, or for purifying a nitrogen-contaminated methane stream by removing nitrogen from the methane stream. The methods can include contacting nitrogenmethane gas mixture with a gas separation membrane, and separating two gas streams at the membrane. One of the separated gas streams would be higher in nitrogen content compared to the nitrogen/methane mixture, and the other separated gas stream would be higher in methane content compared to the nitrogen/methane mixture. In an embodiment of the methods, the gas separation membrane can have a nitrogenmethane selectivity of at least about 5.0 and a nitrogen permeability of at least about 3.0 Barrers. The nitrogenmethane selectivity can be at least about 5.5, at least about 6.0, at least about 6.5, or at least about 7.0. The nitrogen permeability can be at least 3.5 Barrers, at least about 3.9 Barrers, at least about 4.0 Barrers, at least about 5.0 Barrers, at least about 5.5 Barrers, or at least about 6.0 Barrers. In some embodiments, the nitrogen permeability can be at least about 3.5 Barrers and the nitrogenmethane selectivity at least about 6.0, the nitrogen permeability can be at least about 5 Barrers and the nitrogenmethane selectivity at least about 5.5, the nitrogen permeability can be at least about 6 Barrers and the nitrogenmethane selectivity at least about 7.0, or the nitrogen permeability can be at least about 6 Barrers and the nitrogenmethane selectivity at least about 7.5.

The CMS membrane of the disclosure can be used in devices to create a nitrogenmethane separation device, or a gas enrichment device. The membrane can be used in any device in which a CMS membrane could be applied. For example, the CMS membrane could be used as a membrane film 1301 , as shown in FIG. 13 . Gas separation device 1300 includes a nitrogenmethane mixture 1302 , which enters into the device that is spanned by the CMS membrane film 1301 . Nitrogen passes through the film 1301 to produce a nitrogen enriched stream 1303 , and the remaining methane gas which is now depleted of nitrogen exits as a nitrogen depleted methane stream 1304 . As one of ordinary skill in the art would appreciate, separation of the two gases across the membrane film 1301 can be tailored based on flows, pressure, and residence time to produce steams in the desired level of content. For example, the device could be operated in a manner such that stream 1304 is substantially all methane, and the nitrogen enriched stream 1303 includes some amount of methane. Alternatively, the nitrogen stream 1303 could be substantially all nitrogen, and some portion of the nitrogen can exit in the methane stream 1304 . One of ordinary skill would also recognize that the exit streams 1303 or 1304 could be sent to a second separation device for further purification, and a set of films and devices could be set up in series.

The CMS membrane can also be used as a hollow fiber membrane, as shown in FIG. 14 . Hollow fiber membranes of this type have been demonstrated previously for gas separation applications, as shown in U.S. Patent Application Publication No. 2013/0152793, the contents of which are incorporated herein in their entirety. The hollow fiber membrane 1400 shown in FIG. 14 includes a CMS membrane 1401 and a lumen or bore, 1410 , which passes through the length of the hollow fiber. The nitrogen methane mixture 1402 enters the bore or lumen and passes through the length of the hollow fiber. The CMS membrane allows nitrogen to exit through the wall of the hollow fiber to produce the nitrogen enriched stream 1403 while the remaining methane gas passes down the length to exit at the end as a nitrogen depleted methane stream 1404 . The hollow fiber 1400 would be part of a separation device, such that none of the streams 1402 , 1403 , or 1404 would intermix. Multiple hollow fibers can be used in series as a bundle in the separation device.

The description continues in the full USPTO document.